INDEX

Explanations

phrases indicating possession or relationships

oai_token-act-pair · gpt-4o-mini Triggered by @bot

thinking of something

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ItemLayout

-0.45

getAttributes

-0.43

ASM

-0.42

どうやら

-0.40

parently

-0.40

 Cabin

-0.40

HET

-0.39

 مشين

-0.39

 intram

-0.38

iloc

-0.38

POSITIVE LOGITS

CONSIDER

0.60

Think

0.60

consider

0.59

 REMEMBER

0.58

think

0.57

躇

0.57

 Think

0.57

 think

0.56

 Reſ

0.55

remember

0.54

Activations Density 0.008%

phrases indicating possession or relationships

thinking of something

No Comments

No Known Activations